Fat Cells Shrink. Their Number Barely Changes.

What you'll learn

Why the fat cells you carry were largely established while you were still growing

Why belly fat behaves differently from the fat on your hips

How someone at a normal weight can carry metabolically consequential fat around their organs

Why the body pushes back after weight loss — and why regain isn't a failure of will

What liposuction, incretin drugs and metabolic surgery each actually do

How SmartBar applies the deficiency framework

SmartBar's educational framework begins with two organizing terms, and applies them wherever the body has inputs and burdens — starting with what we can verify in food: ingredient composition, level of processing, and nutritional data. Deficiency means something the body needs and isn't getting. Toxicity means a burden the body must process and clear — not a poison.

A word before we start, because this topic gets written about badly. This framework describes biology, not character. Body weight is shaped by genetics, food environment, income, sleep, medication, illness, stress, access to care and personal choices. Reducing it to willpower does not describe the evidence.

This page is educational, not medical advice. It contains no calorie targets, no goal weights and no diet protocol — deliberately. If you are living with or recovering from an eating disorder, some of this material may not be useful to you.

Start with the scale, because the numbers moved faster than most people realise.

In 1990, roughly 731 million adults worldwide were living with overweight or obesity. By 2021 it was 2.11 billion. Among people aged 5–24, 198 million became 493 million. Forecasts put it at 3.8 billion adults by 2050 — more than half the world's adults.[1]

And in 2025, a line crossed for the first time in global estimates. Among children and adolescents aged 5–19, obesity prevalence exceeded underweight — 9.4% against 9.2%.[2] For as long as these records have existed, the problem was too little. For this generation of children, it isn't.

Overweight, obesity, and what BMI can't tell you

For adults, overweight is a BMI of 25 to just under 30. Obesity is 30 or above. For under-18s, separate age- and sex-adjusted criteria apply.[1]

BMI is weight divided by height squared. It says nothing about what the weight is made of, where it sits, or how well your organs are working.

In January 2025, an international commission proposed a better framework. The Lancet Diabetes & Endocrinology Commission recommends using BMI mainly as a population-level risk measure and an initial screening tool, rather than treating it as an individual diagnosis — with excess adiposity confirmed by direct measurement or at least one additional anthropometric measure such as waist circumference, waist-to-hip ratio or waist-to-height ratio.[3]

It distinguishes:

Preclinical obesity — excess body fat with organ function preserved. A state of increased risk, not an illness. The aim is prevention.
Clinical obesity — excess body fat already causing organ dysfunction or limiting daily activities. A chronic disease in its own right. The aim is treatment.

As the Commission puts it, confirming obesity status defines a physical phenotype — it does not by itself represent a disease diagnosis.[3]

Someone above BMI 30 may retain normal organ function, while someone below it may carry clinically important abdominal or organ fat. The framework is new and still being evaluated, but the distinction is the honest one.

Where the fat actually goes

Researchers counted fat cells in 687 adults, dating the cells using carbon-14 left in the atmosphere by nuclear weapons testing. What they found reframes how weight change works.

Adipocyte number remained remarkably stable in the adults studied — including after major weight loss. Twenty people followed through bariatric surgery lost an average of 18 BMI points. Their fat cell number didn't fall. The cells shrank.[4]

The researchers inferred from this stability that adult fat-cell number is largely established during childhood and adolescence, when adipose tissue is still expanding.[4]

There's a myth worth correcting in both directions. It isn't that fat cells never die — roughly 10% are replaced every year, at every age and body size. What stayed stable in these adults was the total.[4]

So weight change is mostly a change in how full the cells are. And on average, obesity can involve more adipocytes, larger adipocytes, or both — participants with obesity here carried up to twice the number of lean participants.[4]

What this means for children. Hundreds of millions are developing obesity while adipose tissue is still expanding. No child's future is biologically fixed, but prevention during growth differs from asking an adult body to reverse an established pattern. Benjamin Franklin's line applies unusually literally here: an ounce of prevention is worth a pound of cure.

Why the body pushes back

Weight-loss advice often leaves out the body's counter-response.

Losing weight does not leave the regulatory system neutral. In a study of 50 adults who lost weight on a ten-week very-low-energy diet, researchers measured appetite hormones at baseline, at ten weeks, and again a year later.

One year on, levels of leptin, ghrelin, peptide YY, cholecystokinin, insulin and several other appetite-related hormones had still not returned to baseline — and neither had hunger. Leptin remained roughly a third below where it started.[5]

The appetite signals encouraging regain had not reverted to baseline. Energy expenditure may also remain lower than predicted from the reduced body size and composition alone — a phenomenon called adaptive thermogenesis.[5]

(Honest limits: the study had high attrition and no weight-stable control group, so it establishes that the adaptations persist, not precisely how much regain each hormone causes.)[5]

Maintaining a lower weight therefore means resisting biological pressure intensified by the loss. Regain after diets, medication or surgery is not evidence that the person stopped caring.

Not all fat is the same fat

Two people can carry identical amounts of fat and face very different risks, depending on where it sits.

Subcutaneous fat sits under the skin — hips, thighs, the layer you can pinch. Visceral fat packs around the organs in the abdomen. Much of the metabolically important visceral fat, particularly omental and mesenteric, drains through the portal circulation directly toward the liver.[6]

That route is one important part of the story: portal drainage can expose the liver to higher local concentrations of free fatty acids and inflammatory signals before they're diluted in the general circulation. The cells' own biology is another — visceral adipocytes show lower adiponectin and leptin, more inflammatory signalling, more lipolysis and weaker insulin response than subcutaneous ones.[6]

Visceral and ectopic fat generally track more strongly with insulin resistance, dyslipidaemia and type 2 diabetes than peripheral subcutaneous fat does.[6]

And fat also accumulates inside organs — the liver especially. Steatotic liver disease associated with metabolic dysfunction also occurs in people without obesity, who show susceptibility to advanced fibrosis and increased all-cause and cardiovascular mortality.[7]

So a person at a "healthy" BMI with fatty liver can carry substantial metabolic risk, while a person at a higher BMI may have a different distribution and a different risk profile. Appearance alone cannot rank the two. Neither can the scale.

The naming changed in 2023. MASLD now describes the metabolic form, MASH its inflammatory progression, and MetALD the overlap with clinically meaningful alcohol intake.[8] That last category recognises that metabolic dysfunction and alcohol exposure can operate together — the old binary definitions forced clinicians to classify people as though only one could be.

Why you can't cut it out

If excess fat is the problem, removing it surgically should help. That has been tested properly.

Researchers performed large-volume abdominal liposuction on 15 women with abdominal obesity — eight with normal glucose tolerance, seven with type 2 diabetes — removing 9 to 10.5 kg of fat, an 18–19% reduction in total body fat.[9]

No significant metabolic improvement was detected over the following 10–12 weeks. Not in insulin sensitivity of liver, muscle or fat tissue. Not in CRP, IL-6, TNF-α or adiponectin. Not in blood pressure, glucose, insulin or lipids.[9]

The explanation ties this whole section together. Liposuction removes subcutaneous fat and reduces fat-cell number — while leaving visceral fat mass unchanged, the remaining cells the same size, and liver and muscle fat untouched.[9] Weight loss achieved through sustained energy deficit commonly reduces visceral and ectopic fat and shrinks the remaining adipocytes — changes that removing subcutaneous tissue alone did not reproduce in this trial.

The small, short trial cannot exclude later or smaller effects — but removing substantial subcutaneous fat did not reproduce the metabolic effects of sustained weight loss.

What excess fat is associated with

Delivered as information rather than as a scare list, because you can only weigh a risk you can see.

Cancer is the one most people don't know about. The International Agency for Research on Cancer concluded there is sufficient evidence that avoiding excess body fatness lowers the risk of at least 13 cancers: oesophageal adenocarcinoma, gastric cardia, colorectal, liver, gallbladder, pancreatic, postmenopausal breast, endometrial, ovarian, renal-cell carcinoma and thyroid cancers, meningioma and multiple myeloma.[10]

The rest, by system: type 2 diabetes, dyslipidaemia and metabolic syndrome; hypertension, coronary disease, stroke, heart failure and venous thromboembolism; MASLD; obstructive sleep apnea; osteoarthritis — including in the non-weight-bearing hand, so this isn't purely mechanical load; chronic kidney disease; PCOS and reduced fertility in both sexes; and substantial bidirectional psychological burden.[11] Our Sleep page covers sleep apnea as a condition many people don't know they have.

The WHO attributes 3.7 million deaths in 2021 to higher-than-optimal BMI.[12]

(One caveat: most of this evidence is observational and does not establish causation for every condition listed. IARC's causal judgement applies specifically to the cancers. A contested "obesity paradox" literature also exists, substantially explained by illness-related weight loss.)

Why "eat less, move more" is true and insufficient

The energy arithmetic isn't wrong. It's a description of the outcome, not an explanation of the cause — like handing someone in debt a calculator.

In a tightly controlled inpatient trial, participants ate about 500 calories a day more, and gained weight, on an ultra-processed diet — even though meals were matched to the unprocessed arm for presented calories, sugar, fat, sodium and fibre.[13]

That trial cannot explain an entire global epidemic by itself. It demonstrates something narrower and important: foods matched on a conventional nutrient panel can still drive substantially different intake under controlled conditions. Products engineered for rapid, easy eating, stripped of the fibre and structure that normally slow you down. Our front-of-package and Nutrition pages take that apart.

People did not collectively lose character. The food environment changed.

The part that gets shamed instead of understood

Weight stigma is often mistaken for motivation. The evidence instead links it to poorer psychological outcomes, disordered eating cognitions and behaviours, and avoidance or delay of healthcare — the opposite of the conditions in which sustainable change becomes easier.[14]

And shame misreads what's happening. For some people eating becomes compulsive — craving overriding intention, the pattern returning despite repeated genuine efforts. Researchers debate whether "food addiction" is the right clinical label; the difficulty of breaking the cycle is not in dispute.

The loop is engineered from both ends. Weight loss triggers biological counter-regulation, as the hormone data above shows, while the food environment is optimised to be overeaten. Then the industry that profits from the outcome sells the solution back — programmes, powders, subscriptions, and increasingly injections.

What helps, starting with free

Protect sleep and its regularity. Short sleep reduces insulin sensitivity and shifts appetite, and a randomized trial found that simply helping people sleep longer reduced measured food intake with no diet prescribed. Our Sleep page has it.

Move, and start below what you think you can do. The benefit curve is front-loaded, and our Exercise page starts with walking and bodyweight work.

Eat food that keeps you full. Protein, fibre, food structure and slower eating act through several satiety pathways, including GLP-1.

Your gut makes GLP-1 too — but not like the drug

Your intestine releases it after eating. Protein, carbohydrate and fat stimulate it through different pathways, while fermentation products from some fibres activate GLP-1-producing cells in experimental models.[15]

That's part of a broader satiety system, not a homemade equivalent of semaglutide. Native GLP-1 is broken down within one to two minutes by the enzyme DPP-4. Semaglutide was engineered to stay active for about a week.[15]

So eat the fibre and the protein — there are good reasons to. But calling any food or supplement "nature's Ozempic" is a category error, and it usually arrives from someone selling it.

What incretin drugs actually do

The evidence is strong, but the drugs differ: semaglutide activates GLP-1 receptors; tirzepatide activates both GIP and GLP-1 receptors.

In STEP 1, 1,961 adults on weekly semaglutide alongside lifestyle support lost a mean 14.9% of body weight at 68 weeks, versus 2.4% on placebo.[16] In SURMOUNT-1, 2,539 adults assigned to tirzepatide lost an average of approximately 15.0% to 20.9% at 72 weeks depending on dose, against 3.1% on placebo.[17] Among the 1,032 participants who began with prediabetes, three-year follow-up found sustained weight reduction and substantially less progression to type 2 diabetes during treatment.[17]

And it isn't only weight. The SELECT trial randomized 17,604 adults with existing cardiovascular disease and overweight or obesity, without diabetes, and found a 20% reduction in major adverse cardiovascular events, with weight loss sustained to four years.[18] Those are hard clinical outcomes.

Withdrawal belongs in the same account. STEP 1 followed a subset for a year after stopping. They had lost 17.3%. One year off treatment they had regained two-thirds of it, and the cardiometabolic improvements had reverted toward baseline for most measures.[19]

That aligns with the counter-regulation described earlier: when treatment stops, much of its opposition to that system stops too. These are better understood as long-term treatments for a chronic condition than as short courses with a predictable end date. Stopping commonly leads to loss of much of the treatment effect, so the likely duration, cost and exit plan belong in the decision from the beginning.

Two further notes. Weight loss from any method includes some lean mass as well as fat, which makes adequate protein, resistance work and attention to strength important during large or rapid losses — especially in older adults. And the pivotal weight-loss trials enrolled predominantly women and predominantly White participants while SELECT was predominantly male, which limits how confidently every estimate generalises.

SmartBar takes no position on whether anyone should take these medications. That is a conversation with a clinician who knows your history.

The treatment liposuction is not

Metabolic surgery is sometimes grouped with liposuction because both occur in an operating room. Unlike cosmetic fat removal, it changes gastrointestinal anatomy, intake and gut signalling. In the Swedish Obese Subjects study — 2,007 surgery patients against 2,040 matched controls on usual care, followed for a median of about 20 years — surgery was associated with lower all-cause mortality (HR 0.77), lower cardiovascular death (HR 0.70), lower cancer death (HR 0.77), and about three additional years of median life expectancy.[20] A separate analysis found it substantially prevented type 2 diabetes.

The honest boundaries, which the authors state themselves. Life expectancy remained 5.5 years shorter than the general population — the gap narrowed, it didn't close. Ninety-day postoperative mortality was 0.2%, and 2.9% needed repeat surgery. The study was prospective and controlled but not randomized, and most procedures used a technique now largely superseded. Surgery also carries nutritional consequences and requires lifelong follow-up.

It is not for everyone, but its decades of outcome evidence cannot be grouped with cosmetic fat removal.

Key Takeaways

Fat loss mostly shrinks adipocytes rather than reducing their total number. In the adults studied, adipocyte number stayed stable even after major weight loss, although individual cells continually died and were replaced.

BMI screens; it doesn't diagnose. A 2025 international commission distinguishes preclinical obesity — excess fat with organs working — from clinical obesity, where function is already affected.

Where fat sits matters more than how much shows. Visceral fat drains toward the liver, and organ fat is invisible from outside.

The body counter-regulates after weight loss. Appetite hormones and hunger were still altered a year later. Regain is biology, not a character verdict.

Treatments are not interchangeable. Liposuction removed substantial subcutaneous fat without measurable short-term metabolic improvement; incretin drugs produce large losses and clinical benefits but commonly require continued treatment; metabolic surgery has decades of outcome evidence alongside operative and nutritional risks.

How to think about body fat in thirty seconds

Ask where it sits, not just what it weighs. A waist measurement adds information about abdominal risk that body weight alone cannot provide.
Looking slim is not a clearance. Organ fat can't be seen from outside, and it's the kind that matters most.
Judge food by how full it keeps you, not only its calorie count — volume, protein, fibre and structure all do work the label doesn't show.
Protect sleep and movement before adjusting anything else. Both change appetite and insulin sensitivity, and both are free.
If you're considering treatment, ask about the exit as well as the entry — what happens when you stop is part of the decision.

What SmartBar does with this

SmartBar scores food. It doesn't score you, weigh you, set you a target, or track your body at all — and on this topic that restraint is deliberate.

What the app does is make one input legible. Scan a product and the scoring engine asks what it gives your body and what your body has to clear: processing assessed first, added sugars and refined starches counted, every deduction shown.

What it can't do is more than that. It can't see your sleep, your stress, your genetics, your income or your neighbourhood — all of which shape this more than most articles admit.

The reason it's free is the argument this page has been making. The people most affected by the food environment are least able to pay for another product promising to fix it. Information shouldn't be the part that costs extra.

This article is educational and is not medical advice. It deliberately contains no calorie targets, goal weights or diet protocols. If you are living with or recovering from an eating disorder, or if reading about body weight is distressing, please speak with a qualified professional rather than acting on anything here. Decisions about weight-loss medication, metabolic surgery, or management of diabetes, liver disease, sleep apnea or cardiovascular risk belong with a clinician who knows your history. If you have been told you have fatty liver, that is worth discussing with your doctor regardless of what you weigh.

Keep exploring: Nutrition · Sleep · Exercise · Lifestyle · Front-of-package claims

Medical Disclaimer: Read our full Medical Disclaimer

References

GBD 2021 Adult BMI Collaborators (2025). "Global, regional, and national prevalence of adult overweight and obesity, 1990–2021, with forecasts to 2050." The Lancet 405(10481):813–838. DOI: 10.1016/S0140-6736(25)00355-1. PMID: 40049186. PMCID: PMC11920007. Companion analysis: GBD 2021 Child and Adolescent BMI Collaborators (2025), The Lancet 405(10481):785–812, DOI: 10.1016/S0140-6736(25)00397-6. (Adults with overweight or obesity: 731 million in 1990 to 2.11 billion in 2021; ages 5–24: 198 million to 493 million; forecast 3.8 billion adults by 2050. Definitions: adults, overweight BMI 25 to <30, obesity ≥30; under-18s per International Obesity Task Force criteria. Authors note universal BMI cutpoints may not capture optimal risk across populations. Forecasts assume continuation of past trends.)
UNICEF (2025). Feeding Profit: How Food Environments Are Failing Children — 2025 Child Nutrition Report; and World Obesity Federation, World Obesity Atlas 2026 — Childhood Obesity. (In 2025, obesity prevalence among children and adolescents aged 5–19 reached 9.4%, exceeding underweight at 9.2% for the first time globally; approximately 188 million aged 5–19 living with obesity, based on data from more than 190 countries.)
Rubino, F., Cummings, D.E., Eckel, R.H., Cohen, R.V., Wilding, J.P.H., Brown, W.A., et al. (2025). "Definition and diagnostic criteria of clinical obesity." The Lancet Diabetes & Endocrinology 13(3):221–262. DOI: 10.1016/S2213-8587(24)00316-4. PMID: 39824205. (Published 14 January 2025. Recommends BMI be used only as a surrogate measure of health risk at population level or for screening, with excess adiposity confirmed by direct fat measurement or at least one anthropometric criterion in addition to BMI; above BMI 40, excess adiposity may be pragmatically assumed. Defines preclinical obesity as excess adiposity with preserved organ function and generally increased future risk, and clinical obesity as excess adiposity with ongoing organ dysfunction or functional limitation. States that confirming obesity status defines a physical phenotype but does not represent a disease diagnosis per se. The framework is recent and implementation and validation remain under evaluation.)
Spalding, K.L., Arner, E., Westermark, P.O., et al. (2008). "Dynamics of fat cell turnover in humans." Nature 453(7196):783–787. DOI: 10.1038/nature06902. PMID: 18454136. (n=687 adults, adipocyte age determined by integration of nuclear-test-derived carbon-14 into genomic DNA. Adipocyte number remained stable in the adults studied, including 20 patients losing a mean ~18 BMI points after bariatric surgery, in whom cell size fell while number did not. Approximately 10% of adipocytes renewed annually at all ages and BMI levels. The inference that number is set during childhood and adolescence follows from observed adult stability; the study did not follow children prospectively into later life, and does not establish that any individual's count is permanently fixed.)
Sumithran, P., Prendergast, L.A., Delbridge, E., Purcell, K., Shulkes, A., Kriketos, A., Proietto, J. (2011). "Long-Term Persistence of Hormonal Adaptations to Weight Loss." New England Journal of Medicine 365(17):1597–1604. DOI: 10.1056/NEJMoa1105816. (n=50 adults with overweight or obesity, without diabetes; 10-week very-low-energy diet, measured at baseline, 10 weeks and 62 weeks. One year on, significant differences from baseline persisted in leptin, ghrelin, peptide YY, cholecystokinin, insulin, gastric inhibitory polypeptide and pancreatic polypeptide, alongside increased subjective hunger; leptin remained approximately 35% below baseline. Limitations: high attrition and no weight-stable control group — the study establishes persistence of the adaptations, not the precise causal contribution of each hormone to regain. Adaptive thermogenesis: Rosenbaum, M., Hirsch, J., Gallagher, D.A., Leibel, R.L. (2008), "Long-term persistence of adaptive thermogenesis in subjects who have maintained a reduced body weight," American Journal of Clinical Nutrition 88(4):906–912, DOI: 10.1093/ajcn/88.4.906.)
Foster, M.T., Pagliassotti, M.J. (2012). "Metabolic alterations following visceral fat removal and expansion: Beyond anatomic location." Adipocyte 1(4):192–199. DOI: 10.4161/adip.21756. PMID: 23700533. (Omental and mesenteric visceral depots drain via the hepatic portal vein while subcutaneous drains into systemic circulation; portal effluent in obesity carries higher concentrations of free fatty acids and interleukin-6. Visceral adipose tissue shows reduced adiponectin and leptin, increased inflammatory cytokines, enhanced lipolysis and reduced insulin response versus subcutaneous. Not all intra-abdominal depots drain identically. Rodent visceral-fat removal and transplantation studies give conflicting results and human findings are contested — the drainage anatomy is well supported; "remove it and metabolism improves" is not.)
van Kleef, L.A., de Knegt, R.J., Brouwer, W.P. (2024). "Metabolic dysfunction associated fatty liver disease in healthy weight individuals." Obesity Reviews. PMCID: PMC11449956. (A review, so the outcome statements derive from observational cohorts synthesised within it rather than from a single prospective comparison. Terminology note: the review largely concerns MAFLD- and NAFLD-era cohorts, and MAFLD is not interchangeable with MASLD — it does not exclude alcohol and uses different criteria. The body therefore refers to steatotic liver disease associated with metabolic dysfunction across evolving definitions. "Healthy-weight" is preferred to "lean" because BMI thresholds vary by ethnicity and region.)
Rinella, M.E., Lazarus, J.V., Ratziu, V., et al. (2023). "A multisociety Delphi consensus statement on new fatty liver disease nomenclature." Journal of Hepatology 79(6):1542–1556. DOI: 10.1016/j.jhep.2023.06.003. PMID: 37364790. (Steatotic liver disease adopted as the overarching term; NAFLD replaced by MASLD, requiring hepatic steatosis plus at least one cardiometabolic risk factor; NASH replaced by MASH; MetALD created for the overlap of metabolic dysfunction and greater-than-minimal alcohol intake, which the previous binary definitions excluded.)
Klein, S., Fontana, L., Young, V.L., Coggan, A.R., Kilo, C., Patterson, B.W., Mohammed, B.S. (2004). "Absence of an Effect of Liposuction on Insulin Action and Risk Factors for Coronary Heart Disease." New England Journal of Medicine 350(25):2549–2557. DOI: 10.1056/NEJMoa033179. PMID: 15201411. (15 women — eight with normal glucose tolerance, seven with type 2 diabetes. Large-volume abdominal liposuction removed 9.1 ± 3.7 kg and 10.5 ± 3.3 kg of fat respectively, an 18–19% reduction in total fat. No significant change in insulin sensitivity of liver, skeletal muscle or adipose tissue by euglycemic-hyperinsulinemic clamp; no significant change in CRP, IL-6, TNF-α, adiponectin, blood pressure, glucose, insulin or lipids over 10–12 weeks. Authors note liposuction reduces subcutaneous fat and total fat-cell number without altering visceral fat mass, remaining fat-cell size, intramyocellular fat or intrahepatic fat. Small sample and short follow-up; published correspondence in NEJM argued these limits could have masked an effect.)
Lauby-Secretan, B., Scoccianti, C., Loomis, D., et al. (2016). "Body Fatness and Cancer — Viewpoint of the IARC Working Group." New England Journal of Medicine 375(8):794–798. IARC Handbooks of Cancer Prevention, Volume 16. (Sufficient evidence that absence of excess body fatness lowers the risk of 13 cancers. This is a cancer-prevention judgement, not a Group 1/2A hazard classification of the kind applied to ultraviolet radiation or night shift work.)
Obesity-associated comorbidities. StatPearls, "Obesity and Comorbid Conditions" (NCBI Bookshelf NBK574535); Endotext, "Clinical Problems Caused by Obesity" (NBK278973). (Largely observational; association does not establish causation for every listed condition.)
World Health Organization, "Noncommunicable diseases: Childhood overweight and obesity" (2025). (Higher-than-optimal BMI estimated to have caused 3.7 million deaths from noncommunicable diseases in 2021.)
Hall, K.D., Ayuketah, A., Brychta, R., et al. (2019). "Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain." Cell Metabolism 30(1):67–77.e3. DOI: 10.1016/j.cmet.2019.05.008. PMID: 31105044. (~508 kcal/day higher intake despite meals matched on presented calories, macronutrients, sugar, sodium and fibre. n=20 — a mechanistic demonstration, not an explanation of global prevalence trends.)
Weight stigma. Jeffers, L., Manner, J., Jepson, R., McAteer, J. (2024). "Healthcare professionals' perceptions and experiences of obesity and overweight and its management in primary care settings: a qualitative systematic review." Primary Health Care Research & Development 25:e5. DOI: 10.1017/S1463423623000683. PMID: 38229563. Levinson, J.A., Kinkel-Ram, S., Myers, B., Hunger, J.M. (2024). "A systematic review of weight stigma and disordered eating cognitions and behaviors." Body Image 48:101678. DOI: 10.1016/j.bodyim.2023.101678. PMID: 38278088. PMCID: PMC11180546. (Associations with reduced healthcare use, biased clinical encounters, disordered eating cognitions and behaviours, and psychological distress. The article claims association, not that stigma has been experimentally shown to prevent weight loss.)
Endogenous GLP-1 secretion. Bodnaruc, A.M., Prud'homme, D., Blanchet, R., Giroux, I. (2016). "Nutritional modulation of endogenous glucagon-like peptide-1 secretion: a review." Nutrition & Metabolism 13:92; Tolhurst, G., Heffron, H., Lam, Y.S., et al. (2012). "Short-chain fatty acids stimulate glucagon-like peptide-1 secretion via the G-protein-coupled receptor FFAR2." Diabetes 61(2):364–371. (GLP-1 released from intestinal L-cells in response to nutrient sensing; protein, carbohydrate and fat stimulate through different pathways; short-chain fatty acids from fibre fermentation activate FFAR2/FFAR3 on L-cells in experimental models. Endogenous GLP-1 half-life approximately 1–2 minutes owing to DPP-4 degradation, versus approximately one week for semaglutide. Acutely raising endogenous GLP-1 does not establish that a food's satiety effect operates primarily through GLP-1 — volume, structure, eating rate, protein and multiple hormones all contribute.)
Wilding, J.P.H., Batterham, R.L., Calanna, S., et al. (2021). "Once-Weekly Semaglutide in Adults with Overweight or Obesity." New England Journal of Medicine 384(11):989–1002. NCT03548935. (n=1,961; mean weight change −14.9% versus −2.4% placebo at week 68. Trial populations approximately 74% female and 75% white. Industry-funded.)
Jastreboff, A.M., Aronne, L.J., Ahmad, N.N., et al. (2022). "Tirzepatide Once Weekly for the Treatment of Obesity." New England Journal of Medicine. DOI: 10.1056/NEJMoa2206038. SURMOUNT-1, NCT04184622. (n=2,539; 72 weeks; mean weight reduction 16.0% at 5 mg to 22.5% at 15 mg versus placebo. Tirzepatide is a dual GIP and GLP-1 receptor agonist, mechanistically distinct from semaglutide. Efficacy-estimand figures of 16.0–22.5% are also reported; the values used in the body are the treatment-regimen estimand. Three-year extension: Jastreboff, A.M., le Roux, C.W., Stefanski, A., et al. (2025), "Tirzepatide for Obesity Treatment and Diabetes Prevention," NEJM 392(10):958–971, DOI: 10.1056/NEJMoa2410819 — among 1,032 participants with prediabetes at baseline, diabetes developed in 1.3% on tirzepatide versus 13.3% on placebo at 176 weeks. Funded by Eli Lilly; the lead investigator discloses trial work, advisory roles and consulting with Eli Lilly, Novo Nordisk, Pfizer, Boehringer Ingelheim, Rhythm Pharmaceuticals and WeightWatchers — disclosed on the principle applied throughout this library. Long-term cardiovascular outcome evidence for tirzepatide is not yet equivalent to semaglutide's.)
Lincoff, A.M., Brown-Frandsen, K., Colhoun, H.M., et al. (2023). "Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes." New England Journal of Medicine 389(24):2221–2232. DOI: 10.1056/NEJMoa2307563. PMID: 37952131. (n=17,604 adults with established cardiovascular disease, BMI ≥27 and no diabetes; primary composite cardiovascular endpoint HR 0.80, 95% CI 0.72–0.90. Prespecified analyses report weight loss sustained to 208 weeks and a lower composite kidney endpoint, HR 0.78. 72.3% male, in contrast to the weight-loss trials.)
Wilding, J.P.H., Batterham, R.L., Davies, M., et al. (2022). "Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension." Diabetes, Obesity and Metabolism 24(8):1553–1564. DOI: 10.1111/dom.14725. PMID: 35441470. (Mean loss 17.3% to week 68; following withdrawal participants regained 11.6 percentage points by week 120 — approximately two-thirds — leaving a net 5.6% loss. Cardiometabolic improvements reverted toward baseline for most variables.)
Carlsson, L.M.S., Sjöholm, K., Jacobson, P., Andersson-Assarsson, J.C., Svensson, P.-A., Taube, M., Carlsson, B., Peltonen, M. (2020). "Life Expectancy after Bariatric Surgery in the Swedish Obese Subjects Study." New England Journal of Medicine 383(16):1535–1543. DOI: 10.1056/NEJMoa2002449. NCT01479452. (2,007 surgery patients, 2,040 controls receiving usual obesity care, plus a general-population reference cohort; median follow-up ~20 years. All-cause mortality HR 0.77 (95% CI 0.68–0.87); cardiovascular death HR 0.70; cancer death HR 0.77. Adjusted median life expectancy 3.0 years longer than usual care (95% CI 1.8–4.2) but 5.5 years shorter than the general population. 90-day postoperative mortality 0.2%; 2.9% underwent repeat surgery. Prospective and controlled but not randomized; most procedures were vertical banded gastroplasty, a technique largely superseded. Diabetes prevention reported separately in Carlsson et al. (2012), NEJM 367:695–704.)

Sourcing note: this article cites peer-reviewed trials, international agency assessments and consensus statements with DOI/PMID where available. Sample sizes, study limitations, the distinction between association and established causation, and the difference between a prevention judgement and a hazard classification are stated in the reference notes rather than left implicit. Industry funding and investigator conflicts are disclosed wherever a cited trial carries them, including for trials whose findings this article reports favourably. This article contains no numeric weight, calorie or diet prescriptions by editorial policy. Cross-link targets (#) resolve once the site structure is live. Reading time is measured, not asserted: body words excluding the reference list and the metadata header, counted at 225 words per minute and rounded to the nearest minute.